`minikube docker` subcommand

[owenthereal]

Steps to reproduce the issue:

This is not an issue but a feature request.

It would be nice to have a minikube docker subcommand that shells out to docker with the command's DOCKER_* env vars set to env vars output by minikube docker-env. This is similar to minikube kubectl which runs kubectl against the minikube cluster.

Currently, in order to not to pollute my current shell, I'm running sh -c 'eval $(minikube docker-env) && docker build ...' to self-contain any docker commands against the minikube cluster. This is not as convenient.

Full output of failed command:

N/A

Full output of minikube start command used, if not already included:

N/A

Optional: Full output of minikube logs command:

N/A

[afbjorklund]

Toyed with the idea, when implementing minikube build (which is pretty much the only thing we use Docker for anyway). But it was disregarded both because "minikube is not a package manager" and that it is easy to install docker (not really true, see #6402)

Another practical problem was that docker is not available for all platforms (no binary for windows anymore), and that we would have to do the same for podman. In the worst case ending up having to build and support moby-engine and podman1-remote ourselves.

See #4868

When running docker-machine, it does have a "config" command matching the "env" command:

• https://docs.docker.com/machine/reference/config/
• https://docs.docker.com/machine/reference/env/

But we thought that your current workaround wasn't too much worse than having such a command...

#4868 (comment)

If ignoring the whole idea of providing the client binary for you, I suppose we could implement a "docker-config" and "podman-config" but it would pollute the commands even further. The actual config is also about to get a whole lot smaller soon (ssh).

See #9229

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[spowelljr]

Closing this for now, if there's more evidence that this should be implemented feel free to comment or reopen issue.